Import from Eudora 5 shows HTML as text

Discussion of bugs in Mozilla Thunderbird
Pail
Posts: 21
Joined: September 26th, 2003, 3:48 pm

Import from Eudora 5 shows HTML as text

Post by Pail »

I have imported all my mail but things that were in HTML format now just have the HTML embedded in the message as text.
cheers, Pail
User avatar
tapetv
Posts: 150
Joined: September 27th, 2003, 3:39 am
Location: The Netherlands
Contact:

Post by tapetv »

I had the same problem before w. Mozilla and this morning I converted my Entourage mail using an Applescript to the Eudora-format and then transferred it to Thunderbird...

Same thing: all headers are:

This is a multi-part message in MIME format.

------=_NextPart_000_000D_01C3810C.F5FFAB70
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

bla bla

isn't there some convertor or add-on to change this ?
'Ignore character coding' in TB-prefs does not change it
Arno Peeters
Tape TV Productions
Pail
Posts: 21
Joined: September 26th, 2003, 3:48 pm

Post by Pail »

If you edit the file containing the message(s) and replace "Content-Type: text/plain" with "Content-Type: html/text" it seems to work. A bit tricky to edit though, it seems to be OK if you replace a line but I wouldn't add a line. I had some messages with no "Content "line so I replaced the "MIME version" line. Not good!

cheers, Pail
User avatar
SoaRex
Posts: 369
Joined: April 20th, 2003, 10:01 pm
Location: Japan, the earth, the solar system

Post by SoaRex »

tapetv wrote:I had the same problem before w. Mozilla and this morning I converted my Entourage mail using an Applescript to the Eudora-format and then transferred it to Thunderbird...

Same thing: all headers are:

This is a multi-part message in MIME format.

------=_NextPart_000_000D_01C3810C.F5FFAB70
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

bla bla

isn't there some convertor or add-on to change this ?
'Ignore character coding' in TB-prefs does not change it

Pail wrote:If you edit the file containing the message(s) and replace "Content-Type: text/plain" with "Content-Type: html/text" it seems to work.

Are there any difference between mail headers of original(Entourage) mail and mai headers of Eudra format(Converted by your ApplScript) and mail headers of Thunderbird? ("Transfered" means "import"?)

If there is no information about "Content-Type: text/html" in Entourage or Eudra format mail header nor mail control file of Entourage/Eudra, I think Thunderbird can do nothing since "No Content-Type header" or "No mime type/subtype in Content-Type header" should be treated as "text/plain".
(Note: "No charset attribute" case is slightly different. Although RFC says "no charset attribute" is defaulted as "US-ASCII", Mozilla/Thunderbird try to autodetect on charset for user's convinience.)
User avatar
tapetv
Posts: 150
Joined: September 27th, 2003, 3:39 am
Location: The Netherlands
Contact:

Post by tapetv »

Well, that's the problem with migrating once very two months or so: you have different type of headers and of course the received mail is send from different clients and platformns. These are some examples:

--Apple-Mail-1--728529259
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
charset=ISO-8859-1;
format=flowed


--MS_Mac_OE_3146308311_7967749_MIME_Part
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit


and the most common:

This is a multi-part message in MIME format.

------=_NextPart_000_000E_01C36C97.463C6310
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

So any 'search and replace' action (which is a good idea btw) would render unexpected results.
Funny is (and I've seen this in Mozilla mail to) is if you forward a message like that, it is converted to flawless html most of the time.
Arno Peeters
Tape TV Productions
User avatar
SoaRex
Posts: 369
Joined: April 20th, 2003, 10:01 pm
Location: Japan, the earth, the solar system

Post by SoaRex »

This is probably lo---nglived Bug 3157 of Mozill Mail&News.
Import (formerly imported) HTML messages from Eudora - seen as plain text, <x-flowed>.
See http://bugzilla.mozilla.org/show_bug.cgi?id=3157
You can find many Eudra & HTML related bug reports in Bugzilla by searching with Product=Mailnews & Status=(UNCONFIRMED/NEW/ASSIGNED/REOPENED) & Summary(contains all of the word/string)="Eudora HTML".

Eudora->Outlook Express->Thunderbird maybe a workaround.
Someone reported that importing Outlook mail by Mozilla had problem but Outlook->Outlook Express->Mozilla worked well.
niralisse
Posts: 12
Joined: July 1st, 2003, 3:55 pm

Post by niralisse »

[quote="SoaRex"]Eudora->Outlook Express->Thunderbird maybe a workaround.
quote]
IIRC the problem with that was your date headers got stripped - not a very useful workaround!
User avatar
Binba
Posts: 8
Joined: September 28th, 2003, 7:58 am

Post by Binba »

Replying to tapetv and SoaRex, Thunderbird should not treat <I>no-content-type</I> as <I>text/plain</I>, but simply add <I>Content-Type: html/text</I> to any mail beginning with <I><x-html></I>!

And I reallly can't get it: This bug is major, as it affects many of those who want to migrate to Thunderbird, plus it seems really easy to fix - just add a content-type header - yet <B>it's been known since FEBRAUARY 1999</B> <a href=http://bugzilla.mozilla.org/show_bug.cgi?id=3157>(3157)</a>, and hasn't been fixed!! Why?!

Maybe someone can jot down in a few minutes a program that'll search through all the *. files in a directory and its subdirectories, and replace all <I><x-html></I> with "<I>Content-Type: html/text</I>(#13)<I><x-html></I>"??
Could be done even in DOS...
User avatar
jfsolem
Posts: 9
Joined: October 1st, 2003, 1:40 pm
Location: Oslo, Norway
Contact:

Post by jfsolem »

I've just posted some info to Mozilla-Bugzilla, bug #3157, but repeat them here FYI.

There is a freeware utility for Mac OS X that will help, see: http://homepage.mac.com/aamann/Eudora_M ... eaner.html .

There are also some perl and python Eudora mailbox conversion scripts that also deals with the problems -- none of them handle all problems, but perhaps someone could get good ideas from looking at the code. The utilities are: eud2mbox ( http://www.jjminer.org/eud2mbox/ - perl script, loses attachments, solves x-flowed etc), CEIConvert ( http://www.timcoston.com/linux/index.php - perl scripts, loses attachments, doesn't handle non-us-ascii on the subject line, doesn't handle x-flowed etc.) and eudora2unix ( http://eudora2unix.sourceforge.net/ - python scripts, doesn't handle non-us-ascii, keeps attachments).
Bernard W. Joseph
Posts: 106
Joined: October 17th, 2003, 12:21 pm
Location: Clinton Twp, MI
Contact:

Eudora and html

Post by Bernard W. Joseph »

I had the same problem. It slowly disappeared as I eventually trashed all the imports from Eudora in the normal course of work.

Actually, I dropped Eudora because I want to send and receive only pure text messages. The message itself may be impure, but that's another subject.

Eudora won't compose pure text, although it will send it. I joined the Eudora forums, read a while, and finally participated. I got mildly flamed for saying that I don't want html, and wanted advice as to how to get rid of it. I also got several honest messages asking why I didn't want to send nor receive html.

As soon as I got T-bird running to 99% of my satisfaction, I switched to it.
Bernard W. Joseph
"I do not suffer fools, nor do I fool sufferers."
User avatar
jfsolem
Posts: 9
Joined: October 1st, 2003, 1:40 pm
Location: Oslo, Norway
Contact:

A perl script for post-processing

Post by jfsolem »

As no one else has come up with a bugfix or tool, I decided to do something about it myself. I've made a perl script to (post-)process a new Mozilla (on win32) mailbox created by importing a Eudora (on win32) mailbox. HTML mail in the resulting mailbox(es) will then be rendered correctly in Mozilla, and the mailbox(es) can also be imported/copied again into Evolution or Kmail or whatever if you want (I imported a mailbox into Evolution myself to test it). You can download the perl script here: http://greenhat.no/download/moz-eudora-html-fix.zip
User avatar
SoaRex
Posts: 369
Joined: April 20th, 2003, 10:01 pm
Location: Japan, the earth, the solar system

Re: A perl script for post-processing

Post by SoaRex »

jfsolem wrote:As no one else has come up with a bugfix or tool, I decided to do something about it myself. I've made a perl script to (post-)process a new Mozilla (on win32) mailbox created by importing a Eudora (on win32) mailbox.
HTML mail in the resulting mailbox(es) will then be rendered correctly in Mozilla, and the mailbox(es) can also be imported/copied again into Evolution or Kmail or whatever if you want (I imported a mailbox into Evolution myself to test it).
You can download the perl script here: http://greenhat.no/download/moz-eudora-html-fix.zip

It's great!
Your script will help many Mozilla/Thunderbird users because there are many DUPE bugs for this problem.
Please add a comment and attatch your script to Bug 3157.

By the way, I was very surprized because the magic is simply adding "Content-Type: text/html" for mail starting with <x-html>.
I hope Mozilla developers will introduce your magic into mail import routine immediately.
User avatar
jfsolem
Posts: 9
Joined: October 1st, 2003, 1:40 pm
Location: Oslo, Norway
Contact:

Re: A perl script for post-processing

Post by jfsolem »

Attachment now uploaded to Bug 3157. The solution was actually proposed in this thread by "Binba". A comparison of Mozilla mailboxes containing mail originally received by Mozilla on one side, and mailboxes that were imported from Eudora on the other, indicated also that this was all there was to it!:-)
rduke
Posts: 12
Joined: December 2nd, 2002, 3:20 pm

x-html fix

Post by rduke »

The previously posted script seems to be for PRE-processing Eudora mailboxes.

If you want to POST-process imported files, this quick perl hack seems to have worked for me:
<code>
$/="\nFrom - ";
while (<>) {
if (/\n\n<x-html>\n/
&& ! /^content-type: /mi) {
s/(\n\n<x-html>\n)/\nContent-Type: text\/html; charset=iso-8859-1\n$1/;
}
print;
}
</code>
User avatar
jfsolem
Posts: 9
Joined: October 1st, 2003, 1:40 pm
Location: Oslo, Norway
Contact:

Re: x-html fix

Post by jfsolem »

rduke wrote:The previously posted script seems to be for PRE-processing Eudora mailboxes.

I actually used my own script for POST-processing, but it can of course be used for PRE-processing instead. I like your short and elegant "quick hack", though!:-) Note, however, that your script will work well only with HTML using the iso-8859-1 character set - mening that HTML mail using UNICODE or a different iso-8859-? character set (e.g. slavic languages using iso-8559-2) will be messed up; mine will set the character set to whatever the original HTML had - and only use iso-8859-1 as a default.
Post Reply